A too-tight timing chain causes undue friction (binding) and wear. You can asily verify this by derailing the rear (drive) chain and spinning the "disconnected" timing ring and cranksets. A tight chain, like tight cones, is a waste of energy.
I don't believe any amount of over tightening will cause a timing chain to break. Bicycle chains don't really "break" per se. Instead, they pull apart; usually where they were joined during installation. Chains can also pull apart from shifting under high pressure; especially with a modern "ram ped" cogset (i.e. hyperglide). Finally, a chain that's been jammed between chainrings and/or against the chainstay might later pull apart as a result of the earlier trauma. The final "normal" cause of spontaneous chain failure is the only one where the metal is fractured: severe lack of lubrication. Any bike shop mechanic will tell you it is unwise to "revive" a rusty chain by oiling it--as soon as you rotate the pedals, some of the frozen outer connecting links will develop radial hairline fractures from the pin hole. Later, after a crack propagates under pedal pressure, the outer link splits open and the chain pulls apart. Except for the above, I can't say that chains really "break." The current dimensions of bicycle chain were developed over 100 years ago and intervening advances in manufacturing and metallurgy have probably made them "unbreakable."
This is not to say that chains don't wear out. We use the word "stretch" but chains don't really stretch any more than they break. What happens is that a chain rapidly gets longer when it's first put into service because each pin re-seats itself. Then, as the pins and bushings wear due to load and friction, the chain will slowly continue to lengthen. Because a chain that no longer has the correct "pitch" (distance between links) causes premature chainring and cog wear, using a worn chain is "penny wise and pound foolish." A recent posting asked "How long does a timing chainring last?" My answer: if you replace your chains when they're worn, you won't need to worry about replacing chainrings. Finally, as if destroying cogs and chainrings isn't bad enough, worn chains are inefficient, noisy and (on the rear of a tandem) impair shifting performance.
The leading causes on chain wear are improper lubrication (insufficient amount or poor quality) and fine grit. I regularly spray on liberal amounts of Tri-flow and then remove grit and excess oil by squeezing a rag around the chain with one hand while turning the pedals backwards with the other. I personally don't believe in removing a chain to clean it--the result is an increased incidence of chain pull-apart. I also don't believe in bathing the chain in solvent--it flushes out the original factory-installed lubrication. Because high priced motorcycle "chain lubes" were developed to stay in place at high r.p.m., they're too sticky for bicycles. I also generally distrust "clean" lubes (usually paraffin based), "environmentally friendly" lubes (repackaged salad oil?), and "penetrating/rust fighting" lubes (which contain non-lubricating solvents and acids). While there are probably dozens of suitable lubes, Tri-Flow works fine for me.
Did I get off-track or what?
Timing chain tension. At the Santana factory, we start with too-tight an initial chain tension so our customers won't have to adjust the chain when it takes its initial "break-in" stretch (within 20 miles). After a timing chain has taken its initial stretch it should probably be oiled and adjust ed every 200-300 miles. What is the proper adjustment? As tight as possible without binding. Because most chainrings aren't perfectly concentric, as you rotate the cranks tight and loose spots are normal. Thus the oft-repeated "measured slack" method of adjustment is not only messy, it's impre cise. Instead, without ever touching the chain I simply adjust the eccentric to where the chain vibrates (like a lightly-plucked guitar string) while backpedaling through the very tightest segment of the chain's rotation. If you want to verify the accuracy of my method, derail the rear chain to check for binding.
On newer Santana tandems we've gotten the eccentric and housing tolerances exact enough to allow the following "near-automatic" adjusting sequence (requires no trial-and-error, takes less than 60 seconds). First, rotate the cranks to the loosest part of the rotation (where the chain droops the most). After loosening the two eccentric grub screws one full turn each, use a pin-wrench to rotate the eccentric and snug the chain. Finally, backpedal a half dozen complete revolutions and then retighten the two grub screws. Because a too-tight chain will rotate a loose-but-not-wobbly eccentric, backpedaling will automatically rotate the eccentric to its optimal position.
Warning to those who contend it's too much trouble to tighten a sagging timing chain: A too-loose chain can not only jump all the way off, much worse is when it jams while coming only part way off. This will generally happen during a stand-up sprint when frame bowing creates just enough additional chain slack to allow the timing chain to mistrack by a single link. This is an experience the two of you will never forget--when the drivetrain locks-up, your spin rate instantaneously drops to zero. If you're lucky you'll only fold the chainring, dump the chain and stay upright. If you're extremely unlucky you'll "bust" both knees before you hit the ground. If you think this can't happen, here's an experiment to convince disbelievers: loosen the timing chain just enough to allow you to dislodge a single link at the bottom of the captain's timing ring. With the single link dislodged, tighten the eccentric. Now try to backpedal--you can't--even if you jump up and down on the crank (if you could jump hard enough you'd break a chainring bolt, fold the ring and dump the chain). Of course when you're sprinting instead of backpedaling, the chain will mistrack and jam at the bottom of the stoker's chainring.
Pin tools. The cheapest, smallest, lightest and strongest pin wrenches are made by Park Tool and can be purchased at any good bike shop for about $10. The red or green handled versions both fit a Santana eccentric.
If your frame was properly designed, tightening the eccentric won't cause you to need to remove a link. If you do find yourself removing links on fairly-new timing chains, you can correct the original design error by changing both timing rings from an even to odd number of teeth or vice-versa. Further, if the builder has done his homework, normal readjustment occurs through the high or low arc of the eccentric and seat height is affected to a minimal degree. Finally, Santana tandems are designed to have the front spindle rotated through the lower half of the rotation--using the upper half raises the captain's center of gravity and decreases stability.
Bill McCready; Design Dude
Santana Cycles, Inc.
P.S. After composing this entry, I read with interest the posting from a Santana owner who attributed bottom bracket failures and drivetrain jamming to small timing chainrings. When he replaced the original chainrings with a larger pair the problem ceased. If he still has the original rings, I' m willing to bet a jar of Tang that the stoker chainring had one or more bent teeth. I'll also predict a close inspection of the same ring will reveal gashes from where the timing chain mis-tracked. A timing ring with one or more bent teeth will facilitate the same sort of mistrack-and-jam probl em as a too-loose chain. In his case the mistracking and chain-jam was not as traumatic because the bent chainring allowed the mistracking to occur while seated (instead of sprinting). Still, if my theory is correct, repeated instances of chain-lock would certainly have lead to premature bearing failures. Changing to larger new chainrings solved his problem. I am guessing that straightening one or more bent teeth on the existing smaller rings would have been just as effective.